maiImage / createImage

Here's a node that is almost aggressively minimal. maiImage / createImage takes exactly one required input - prompt - plus an aspect ratio, a quantity, and an output format, and that's the whole story. No negative prompt, no seed, no steps, no cfg, no sampler, no LoRA socket. It's a pure "describe it, get an image" node, the kind you'd expect from a hosted product API rather than a ComfyUI sampler, which is exactly what it is: MaiImage is one of the engines Civitai's orchestration reaches through fal, and this node is a thin, generated wrapper around that API's createImage operation.

Full honesty: there is almost no public documentation on what "MaiImage" actually is. It shows up in Civitai's orchestration spec (MaiImageCreateFalImageGenInput), it's billed in Buzz like everything else in this pack, and it returns a standard IMAGE output - but there's no model card, no open-weights repo, and no meaningful community discussion to point you at. That's not a reason to distrust it, but it is a reason to test cheap before you commit. If you're the kind of person who needs to know exactly which model is rendering your image before you spend money on it, this node will frustrate you.

The inputs that matter

• prompt - the only required input, and it's doing all the work. There's no negative prompt to lean on, so positive prompt craft matters more than usual here.
• aspect_ratio - defaults to auto, with eleven ratio choices (portrait, landscape, square, and everything between). The auto default is a nice touch: it lets the model pick a ratio that fits your prompt, which is genuinely handy for an API with no width/height fields.
• quantity - how many images per call.
• output_format (jpeg/png/webP) and image_metadata (external metadata stored with the output) - the pack standards.

Outputs: images (IMAGE), errors, workflow_id, raw_json. The node shows the job id and per-job cost inline - the pack's way of keeping cloud billing honest.

Install and auth

The one install covers the whole pack. ComfyUI Manager → Civitai Comfy Nodes, or:

cd ComfyUI/custom_nodes
git clone https://github.com/civitai/civitai-comfy-nodes.git
pip install -r civitai-comfy-nodes/requirements.txt

Authenticate with a Civitai account that has Buzz: export CIVITAI_API_TOKEN=..., a Civitai Auth node, or the sidebar sign-in.

The honest call

If you found this node while browsing the Civitai menu and wondering what it does - same. The sensible move is to run one draft, look at the output, and decide for yourself whether MaiImage earns a place in your workflow. There are better-documented siblings in the pack for when you want control (Krea 2 comfy base, Flux2-dev, Qwen) and this node for when you want to type a sentence and get an image with zero knobs. On a no-GPU machine that's a real use case. Just don't build a production workflow on an engine this undocumented without seeing its output first - and don't blame the node for being thin; it's a generated wrapper, and it does exactly what it says.